Abstract
Objectives: To evaluate the outcomes of adjuvant radiotherapy in patients with esophageal cancer (EC) who underwent esophagectomy following neoadjuvant chemoradiotherapy (NCRT).
Methods: The data of EC patients who received adjuvant therapy after NCRT between 2004 to 2019 was retrieved from the SEER database. The patients were split into the adjuvant radiotherapy with or without chemotherapy (RT±CT) and the adjuvant chemotherapy (CT) groups. The process of propensity score matching (PSM) was employed.
Results: Following PSM, 157 patients in total were recruited in each treatment group. There were no significant variations in either overall survival (OS) or cancer-specific survival (CSS) between the RT±CT and CT groups (median OS: 28 months versus. 51 months, p=0.063; median CSS: 31 months versus. 52 months, p=0.16). Within the CT group, patients with ypI/II or cI/II tumor stage, positive lymph node ratio (LNR) ≤0.1, and tumor size ≥50 mm (p<0.05) had higher OS compared to the RT±CT groups. Among patients with cT3-4 tumors in N-stage downstaging group, the OS and CSS were significantly greater for those underwent RT±CT as opposed to the CT group (5-year OS:56.6% versus 19.4%, p=0.042; 5-year CSS:67.9% versus. 19.4%, p=0.023). Multivariate Cox regression analysis identified the tumor histology grade as an independent prognostic factor of OS and CSS.
Conclusion: Radiotherapy-based adjuvant therapy does not significantly improve the prognosis of EC patients after NCRT, although it may provide a survival benefit for patients with cT3-4 tumors in N-stage downstaging.
Previous research demonstrated that patients with resectable esophageal cancer (EC) can experience considerably higher survival rates while receiving neoadjuvant chemoradiotherapy (NCRT) in addition to esophagectomy.1 Therefore, NCRT is recommended as the standard treatment modality for patients with resectable EC in the NCCN guidelines (version 2.2023).2 Nevertheless, the survival chances after surgical resection following NCRT are still dismal, with distant metastases and local recurrence being the primary causes of failure.3-5 Hence, there is an urgent need for a more effective postoperative management model to enhance the long-term survival prognosis of patients. While some research indicates that postoperative adjuvant therapy boosts survival rates for EC patients undergoing NCRT, there is inefficient comparison of postoperative adjuvant treatment modes.6-8 The role of postoperative adjuvant radiotherapy in the management paradigm remains unclear. Hence, this research aims to assess the outcomes of postoperative adjuvant radiotherapy in EC patients after NCRT using SEER database.
Methods
The clinical information of individuals diagnosed with EC between 2004 and 2019 were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. The permission to access the SEER Research Plus database was obtained on Nov 2021 Sub (2000–2019). This study was conducted in The Fourth Hospital of Hebei Medical University, Hebei, China. Previous studies related to this project can be searched through PubMed and Web of science.
The inclusion criteria were i) The pathological diagnosis indicated primary esophageal cancer; ii) received preoperative chemoradiotherapy; iii) received esophagectomy; iv) received adjuvant therapy. The exclusion criteria were i) staging at M1/MX stage and TX/NX stage (AJCC 8th); ii) Unknown preoperative and postoperative treatments.
The screening process is shown in Figure 1. The adjuvant chemotherapy (CT) group and the adjuvant radiotherapy with or without chemotherapy (RT±CT) group comprised the included patients. The study used established data and did not involve interaction with patients. Approval has been granted by our Institutional Review Board, and this research strictly adhered to the principles outlined in the Helsinki Declaration.
- Inclusion and exclusion flow diagram for SEER EC patients who received adjuvant therapy after esophagectomy following NCRT between 2004 to 2019. EC: sophageal cancer, NCRT: neoadjuvant chemoradiotherapy
Statistical analysis
The Chi-square test was employed to compare the baseline characteristics between the RT±CT and CT groups. Propensity score matching (1:1) was performed for age, tumor histology, histological grade, tumor location, T stage (yp/c), number of lymph nodes (LNs) removed, and count of pathologically positive LNs (ypN) using the Stata software (version 16.0). The Kaplan-Meier method was employed to estimate overall survival (OS) and cancer-specific survival (CSS), and the log-rank test was performed to compare survival curves. Forest plots showed the results of subgroup analysis. Ultimately, variables associated with OS and CSS after PSM were identified using univariate and multivariate Cox regression models. The multivariate analysis included variables that showed a p-value of 0.2 or less in the univariate model. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, N.Y., USA). The forest plots and survival curves were plotted using the “survminer” and “forestplot” packages of R statistical software (version 4.2.2). A 2-sided p-value <0.05 was deemed statistically significant.
Results
An entire group of 489 patients were included for this analysis. After PSM, 314 patients were analyzed and classified into the RT±CM group (n=157) and the CT group (n=157). The clinicopathological features of the 2 groups did not differ significantly, according to the results of the Chi-square test. Table 1 displays the study population’s baseline characteristics. Survival analysis before PSM.
- Baseline characteristics of patients included in the analysis before and after propensity score matching (PSM).
Survival analysis before PSM
Before PSM, the group as a whole had median OS (mOS) and CSS durations of 34 months (95% Cl, 30-43 m) and 38 months (95% Cl, 31-49 m) before PSM. In addition, the 3-year OS and CSS rates were 48.6% and 50.4%, and the 5-year OS and CSS rates were 35.5% and 37.9%. The mOS duration was 42 months (95% Cl, 32-54m) in the CT group and 28 months (95% Cl, 23-35m) in the RT±CT group, and the difference was not significant (p=0.096). Likewise, the median CSS (mCSS) durations in the CT and RT±CT groups were also similar (p=0.22, Figure 2A) at 44 months (95% Cl, 34-54m) and 31 months (95% Cl, 25-41m). The 3-year OS (54% vs. 38.9%, p=0.096) and CSS (55% versus. 41.9%, p=0.22) rates were greater in the CT group, whereas the 5-year or long-term survival rates were similar in both groups (5-year OS:36.9% versus32.3%, p=0.096; 5-year CSS:37.9% versus 38%, p=0.22).
- Kaplan-Meier curves of overall survival (OS) and cancer-specific survival (CSS) for patients who underwent CT and RT±CT. A) Curve showed the CSS for the entire cohort before PSM. B) Curve showed the CSS for the entire cohort after PSM. C & D) Survival curves showed the OS and CSS for the patients with cT3-4 stage. E & F) Survival curves showed the OS and CSS for patients with N-stage downstaging group (ypN stage is lower than the cN stage). G and H survival curves showed the OS and CSS for patients with cT3-4 tumors in N-stage downstaging.
Survival analysis after PSM
After PSM, the CT and RT±CT groups had mOS durations of 51 months (95% Cl, 31-58 m) and 28 months (95% Cl, 23-35 m) (p=0.063), and the mCSS durations were 31 months (95% Cl, 25-41 m) and 52 months (95% Cl, 32-77 m) (p=0.16, Figure 2B). In contrast to the RT±CT group, the 3-year OS (56% versus 38.9%, p=0.063) and CSS (56.4% versus 41.9%, p=0.16) rates of the CT group were higher. In contrast, the 5-year OS (39.7% versus 32.3%, p=0.063) and CSS (40.7% versus 38%, p=0.16) rates were similar for both groups.
We next evaluated the survival of the 2 treatment groups across the different clinical subgroups, including overall TNM stage, ypN stage, T stage (yp/c), number of nodes excised, tumor size, and positive lymph node ratio (LNR). The OS of the CT group was considerably superior than that of the RT± CT group among patients with ypI/II stage tumors (mOS: 54 m versus 24 m, p=0.008), cI/II stage tumors (mOS: 124 m versus 27 m, p=0.033), tumor size ≥50 mm (mOS: 54 m versus 30 m, p=0.023) and LNR ≤0.1 (mOS: 56 m versus 33 m, p=0.035). Furthermore, the CSS was also significantly higher for the CT group among the individuals with stage ypI/II tumors (mCSS: 54 m versus 24 m, p=0.005). The outcome hazard ratios for CSS are shown in Figure 3.
- Forest plot for esophageal cancer patients in the subgroup analysis (CT vs RT±CT). Hazard ratio (HR) with 95% confidence interval (Cl) for death in terms of the cancer-specific survival (CSS) of patients with esophageal cancer who underwent CT or RT±CT. P-values of the Cox proportional hazards regression are reported. The left side of the invalid line indicates that CT is better and the right side of the invalid line indicates that RT±CT is better.
Moreover, Kaplan–Meier analysis stratified by the ypT stage, ypN stage, and the number of LNs removed revealed a significant variance in the OS among the two groups only in the ypN0-1 subgroup (mOS: 56 m versus 31 m, p=0.021), indicating that patients receiving CT had an obvious survival advantage. In contrast, RT±CT did not confer any survival benefit to patients with ypT3-4 stage tumors (mOS: 41 m versus 23 m, p=0.13; mCSS: 51 m versus 24 m, p=0.2) and removal of fewer than 15 lymph nodes (mOS: 38 m versus 25 m, p=0.081; mCSS: 38 m versus 25 m, p=0.19).
Based on the initial N stage (cN) and the ypN stage, we further divided the patients into the N-stage non-downstaging group (ypN stage consistent with cN stage) and the N-stage downstaging group (ypN stage lower than the cN stage). In the downstaging group, the RT±CT group had marginally higher survival compared to the CT group (p=0.54, Figure 2E). Moreover, the RT±CT group had somewhat improved 3-year OS (p=0.54, Figure 2E) and CSS (p=0.3, Figure 2F) rates. Furthermore, patients in the RT±CT group who were initially staged as cT3-4 showed a slight improvement in 5-year or long-term OS (p=0.423, Figure 2C) and CSS (p=0.734, Figure 2D). The OS (p=0.0442, Figure 2G) and CSS (p=0.023, Figure 2H) rates were significantly better for cT3-4 stage patients in the RT±CT group with N-stage downstaging, whereas no survival benefit was observed in the absence of N-stage downstaging (5-year OS: 28.5% versus 38.4%, p=0.063; 5-year CSS: 31.2% versus 38.4%, p=0.12).
Univariate and multivariate analyses
Univariate analysis revealed that histological tumor grade, total TNM stage, T stage (yp/c), ypN stage, and LNR were all significantly correlated with OS and CSS (p<0.05). Additionally, the tumor histology grade was found to be an independent predictive predictor for both OS and CSS using multivariate Cox regression analysis. Poor/undifferentiated tumors and ypN3 stage were independent predictors of worse CSS (Figure 4).
- Univariate and multivariate Cox regression analysis of factors affecting Cancer-specific survival (CSS). (*unknown subgroups were not included in the analysis)
Discussion
As recommended by the NCCN guidelines (version 2.2023), there is no definitive agreement on the utilization of a radiotherapy-based adjuvant approach following NCRT or NAC for EC patients, regardless of margin status. Recently, multiple studies have indicated that postoperative adjuvant chemoradiotherapy can enhance the survival of some EC patients after NCRT.6-8 However, the role of postoperative radiotherapy in them was not further explored in these studies. Therefore, we compared the OS and CSS of EC patients who received RT±CT or CT as postoperative adjuvant therapy after NCRT using data retrieved from the SEER database. Our findings indicate that RT±CT did not confer any significant survival benefit to the patients compared to CT (p>0.05). This may be attributed to the adverse effects of re-radiation, such as radiation-induced esophagitis or pneumonia, or the hematological complications, which can affect patient survival. In contrast, chemotherapy has a relatively good safety and tolerability profile.9-13 Furthermore, the radiation dose is also a significant factor in the translation of the local effects of radiotherapy to long-term survival.
Previous researches have demonstrated that postoperative adjuvant chemoradiotherapy can effectively improve the survival rate of EC patients whose tumors are at theypT3-4 stage or persistent positive lymph nodes (ypN+) after NCRT. However, the control arm in these studies consisted of patients who did not receive postoperative adjuvant therapy.7,8 Therefore, we also analyzed the ypN+ or ypT3-4 subgroups in our study, and found that there was no survival benefit of RT±CT versus CT in the ypN+ patients, especially those staged as ypN1. This is similar to previous reports that showed a survival benefit of adjuvant chemotherapy but not of adjuvant radiotherapy in patients with ypN+ following NCRT.14,15 We hypothesize that the insignificant improvement in patient survival after adjuvant radiotherapy following NCRT may be the result of increased re-radiation-induced toxicity. However, there are few studies have focused on reducing the toxicity of re-irradiation after NCRT, and only some have mentioned the issue of radiation dose. In one study, conventional fractionated irradiation with 20–30Gy selected on the basis of the dose and risk-tolerant organs of NCRT resulted in better tolerance.7 Nevertheless, another study showed that ≥45Gy conventional fractionated irradiation did not improve patient survival, although the toxicity was not recorded.14 Therefore, the dose of re-irradiation warrants further consideration. Unfortunately, since the SEER database does not contain information regarding the radiation dose and toxicity, we were not able to validate the impact of the radiation dose on patient survival. Furthermore, the OS and CSS were worse in the RT±CT group compared to the CT group for patients with ypT3-4 tumors. We hypothesize that some patients in this group did not respond well to chemoradiotherapy and did not have a significant reduction in the T stage following NCRT, which likely resulted in low sensitivity to re-irradiation, and eventually shorter survival duration. Interestingly, RT±CT improved the survival of patients with N-stage downstaging (Figure 2), indicating that these patients had superior response to chemoradiotherapy, and the therapeutic advantages of re-irradiation likely outweighed its toxic side effects. Therefore, the response of patients to NCRT may affect the result of postoperative adjuvant radiotherapy.
Some studies have shown that even though patients with the initial cT3-4 stage can achieve complete pathological remission (pCR) following NCRT, over 30% of the patients still experience recurrence after surgery, and the initial cT3-4 stage is still an independent risk factor for tumor recurrence.16 However, fewer studies have demonstrated that EC patients at the cT3-4 stage require postoperative adjuvant therapy following NCRT. We observed a trend towards 5-year and longer survival rates of patients with cT3-4 stage in the RT±CT group, although there was no statistically significant distinction among the RT±CT and CT groups (Figure 2). Moreover, patients at the cT3-4 with N-stage downstaging had a significantly higher survival rate in the RT±CT group. In the absence of N-stage downstaging, however, patients receiving RT±CT did not have a survival advantage. This indicated that patients who achieve N-stage downstaging may benefit from postoperative radiotherapy. We were unable to ascertain whether T-stage downstaging was achieved after NCRT in this subgroup due to database restrictions. Nevertheless, we hypothesize that re-irradiation may improve survival in patients with T/N stage downstaging.
The subgroup analysis showed that patients with stage cI/II and ypI/II tumors did not benefit from postoperative adjuvant radiation in terms of OS (p<0.05). Specifically, most of these patients had early-stage cancer without lymph node metastases (T2-3N0M0), and were therefore at a lower risk of recurrence.17 Re-radiotherapy may cause more detrimental side effects, and may be less beneficial to patient survival. Second, according to the findings of the CROSS trial, patients who underwent NCRT and esophagectomy had a higher likelihood of developing distant metastases rather than localized isolated recurrence, suggesting that local radiotherapy may offer limited advantages.4,18 For patients in stage cI/II, there remains controversy regarding the necessity to intensify postoperative adjuvant therapy. Given the data from this study, it appears that these patients often present with unfavorable factors that can impact postoperative survival, including cT3 stage, tumor length exceeding 50 mm, and low pathological differentiation grade. We hypothesize that these factors could be significant drivers for considering adjuvant therapy in patients at stage cI/II. Furthermore, despite tumor length being a crucial prognostic factor for EC patients with cT1-2 stage, and the fact that longer tumor lengths correlate with a heightened risk of recurrence and distant metastasis.19-21 Unfortunately, RT±CT did not enhance OS or CSS in these patients. Hence, the selection of adjuvant therapy needs to be considered in combination with the risk of postoperative recurrence and treatment tolerance.
The number of positive lymph nodes after NCRT significantly affects patient prognosis.22 However, it is unclear how many lymph nodes should be removed from the patients after NCRT.23-25 Following the recommendations of a review, we conducted a subgroup analysis with removal of at least 15 LNs as the cut-off.26 However, adjuvant radiotherapy did not confer a survival benefit in either subgroup. This may be related to the increase in adverse effects after re-radiation. At the same time, neither the univariate nor multivariate analyses showed a correlation between the number of LNs and the prognosis (p>0.05). According to a study that evaluated the outcomes of LN dissection for rectal cancer following NCRT, this may be connected to tumor regression grade after NCRT, and may decrease the number of lymph nodes that need to be removed if the regression grade is favorable.27 However, another study showed that lymph node removal should be performed for EC patients regardless of the response to NCRT.28,29 Further randomized clinical trials are required to address this issue. Some studies have shown that the LNR is more helpful in predicting OS of patients with esophageal cancer following NCRT, and high LNR (>0.1) is a risk factor for recurrence.30,31 This is consistent with the findings of the univariate analysis in this study. According to the subgroup analysis, we found that postoperative adjuvant radiotherapy failed to enhance the OS in patients with LNR ≤0.1, which is not surprising given the minimal risk of recurrence in these patients.
Study limitations
It is noteworthy that the predominant histological type (79.3%) in our cohort was esophageal adenocarcinoma (EAC). However, patients with adenocarcinoma, have a lower probability of local recurrence and a higher rate of hematogenous metastasis when compared to those with squamous cell carcinoma (ESCC).32 Thus, systemic therapy may be more successful than local radiation in EAC patients. Furthermore, our findings may not pertain to patients with tumors in the upper or middle third part of the esophagus since the predominant tumor location in our cohort was the lower third of the esophagus (83.7%). More rigorous clinical trials are needed to resolve this issue. As this study included some cases with close or positive surgical margin status, it may not be entirely suitable for patients with R0 resection and needs to be considered with caution by investigators. Additionally, as the study was retrospective, selection bias was inevitable even with the use of PSM to reduce confounding bias. Second, information about radiotherapy target areas, doses, equipment, and chemotherapy regimens is not included in the SEER database. Finally, prognostic factors such as the interval between NCRT and surgery, the condition of the postoperative margins, and the degree of pathological regression following NCRT were also not evaluated. All of these concerns may affect the reliability of our results.
Implications of findings for future research
This article explores the relevance of postoperative adjuvant therapy following NCRT in EC patients. Our findings indicate a potential benefit of postoperative adjuvant radiotherapy, particularly for patients getting cT3-4 tumors in N-stage downstaging. This suggests that individuals presenting adverse prognostic factors and experiencing N-stage downstaging after neoadjuvant therapy might derive greater advantage from postoperative adjuvant therapy. However, the decision to opt for postoperative adjuvant radiotherapy must be tailored to each patient, considering factors like radiotherapy dose, the scope of the radiotherapy target area, and the chemotherapy regimen. These aspects merit further examination and deliberation. We expect the insights gained from this study to provide a significant reference for determining surgical adjuvant treatments following neoadjuvant immunotherapy.
In conclusion, radiotherapy-based adjuvant therapy does not significantly improve the prognosis of EC patients after NCRT, although it may provide a survival benefit for patients with cT3-4 tumors in N-stage downstaging.
Acknowledgment
The authors gratefully acknowledge TopEdit (www.topeditsci.com) for the English language editing of this manuscript.
Footnotes
Disclosure. Authors have no conflict of interests, and the work was not supported or funded by any drug company.
- Received February 3, 2024.
- Accepted July 11, 2024.
- Copyright: © Saudi Medical Journal
This is an Open Access journal and articles published are distributed under the terms of the Creative Commons Attribution-NonCommercial License (CC BY-NC). Readers may copy, distribute, and display the work for non-commercial purposes with the proper citation of the original work.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.